High Availability and Scaling

Question 1

TRUE or FALSE?
Launch configurations support versioning.

Answer 1

FALSE
They are immutable.
Only launch templates support versioning

Question 2

TRUE or FALSE?
Launch templates can be used to configure all autoscaling features.

Answer 2

TRUE
It’s launch configurations that only support some.

Question 3

What are 4 things you can specify in a launch template?

Answer 3

• AMI
• Instance type
• Security group
• Networking info

Question 4

What are the three autoscaling restriction settings?

Answer 4

Minumum, maximum, desired.

Question 5

What’s one example for when would you use a scaling-out lifecycle hook?

Answer 5

For installing software.

i.e. you don’t want the instance to be in service before the software is properly installed.

Question 6

What’s one example for when would you use a scaling-in lifecycle hook?

Answer 6

For capturing application logs.

i.e. so they are not lost when the instance shuts down.

Question 7

What is a lifecycle hook?

Answer 7

A feature that allows you to put instances into a wait state, allowing you to perform tasks before the instance is put into service or before it is fully terminated.

Question 8

TRUE or FALSE?
The autoscaling group config contains the VPC and subnets where the instances will live.

Answer 8

TRUE

Question 9

What’s the difference between simple scaling and step scaling?

Answer 9

Step scaling provides more granular control.

It allows you to define multiple actions to take based on a range of metric values, rather than a single threshold.

Question 10

What could happen if you don’t have a warm-up period in place?

Answer 10

You would get high instance churn, as instances are not ready, fail the health check, and get terminated.

Question 11

What does a CoolDown period do?

Answer 11

It pauses autoscaling for a certain amount of time, to give the system time to adjust.

Helps prevent runaway scaling events.

Question 12

TRUE or FALSE?
You can use Reserved Instances as the minimum count in an autoscaling group to save money.

Answer 12

TRUE

Question 13

If an exam questions mentions an RDS database solution but with unpredictable workloads, think of…

Answer 13

Aurora Serverless

Question 14

TRUE or FALSE?
RDS storage only scales up, not down.

Answer 14

TRUE

Question 15

In the exam, a viable scaling option for a relational database is refactoring and…

Answer 15

changing to DynamoDB

Question 16

TRUE or FALSE?
Standby RDS instances can be used for reads.

Answer 16

FALSE

Question 17

What are the two scaling options for DynamoDB?

Answer 17

Provisioned and On-Demand

Question 18

What is the most cost-effective scaling option for DynamoDB?

Answer 18

Provisioned

Question 19

What is the best scaling option for DynamoDB with an unpredictable workload?

Answer 19

On-Demand

Question 20

TRUE or FALSE?
With DynamoDB, you cannot switch between scaling options.

Answer 20

FALSE

You can, but only twice per 24 hours per table

Question 21

If EC2 instances in an autoscaling group are taking too long to provision, consider using…

Answer 21

Baked AMIs

Question 22

TRUE or FALSE?
Baked AMIs are better than using user data.

Answer 22

TRUE

Question 23

For questions that ask for a highly available solution for a legacy resource where there is only one instance, think of…

Answer 23

Steady state groups

i.e. the min, max, and desired are all set to 1

Question 24

To make sure EC2 instances in an autoscaling group are terminated and replaced when unhealthy, be sure to enable…

Answer 24

health checks on the load balancer.

Question 25

What are the four different types of disaster recovery strategies?

Answer 25

• Backup and Restore
• Pilot Light
• Warm Standby
• Active/Active Failover

Question 26

What is the default CoolDown period of an autoscaling policy?

Answer 26

300 seconds